Boosting the efficiency of Byzantine-tolerant reliable communication

Reliable communication is a fundamental primitive in distributed systems prone to Byzantine (i.e. arbitrary, and possibly malicious) failures to guarantee integrity, delivery and authorship of messages exchanged between processes. Its practical adoption strongly depends on the system assumptions. One of the most general (and hence versatile) such hypothesis assumes a set of processes interconnected through an unknown communication network of reliable and authenticated links, and an upper bound on the number of Byzantine faulty processes that may be present in the system, known to all participants. To this date, implementing a reliable communication service in such an environment may be expensive, both in terms of message complexity and computational complexity, unless the topology of the network is known. The target of this work is to combine the Byzantine fault-tolerant topology reconstruction with a reliable communication primitive, aiming to boost the efficiency of the reliable communication service component after an initial (expensive) phase where the topology is partially reconstructed. We characterize the sets of assumptions that make our objective achievable, and we propose a solution that, after an initialization phase, guarantees reliable communication with optimal message complexity and optimal delivery complexity.